• esri

A Sense Of (Exactly) Where You Are

How mapping big data to establish “ground truth” will save cities time, money, and pedestrians’ lives.

A Sense Of (Exactly) Where You Are

It’s a rainy Monday morning in Brooklyn. A city bus trundles down DeKalb Avenue, flanked by commuters streaming toward the subway station. A hurried pedestrian decides to cross mid-block, but a double-parked box truck is blocking her view of the street and, crucially, the bus driver’s view of her. But in an instant, one of four Mobileye sensors mounted on the bus spots the woman. The dashboard flashes an alarm and sounds an alert, warning the driver of an imminent collision. He brakes, she pauses, and the moment passes—just one in a thousand such occasions that might have ended tragically without an assist from machine vision.


Mobileye, an Israeli startup now owned by Intel, estimates that its Shield+ system could reduce potential collisions in New York, Barcelona, and Tel Aviv by as much as 80%. It’s a significant reduction, especially stateside where traffic fatalities have been steadily on the rise over the past decade. More than 40,000 people in the U.S. died in traffic accidents in 2016, including almost 6,000 pedestrians. Several dozen cities, including New York, have adopted “Vision Zero” initiatives (i.e., zero traffic fatalities or serious injuries) through enforcing lower speed limits, redesigning streets, and changing drivers’ behavior—all of which new technologies can help.

But it’s one thing to save a life and have the lesson vanish; it’s something else altogether to log a near miss, map it, and identify the most dangerous intersections in New York, giving the MTA actionable information. In November, the spatial analytics firm Esri announced it would work with Mobileye to do just that. By streaming Shield+ real-time incident and location data into Esri’s ArcGIS mapping platform, the companies are enabling cities to optimize themselves for safety.

In mapping circles, “ground truth” refers to observations collected firsthand. In the smartphone era, companies such as Uber and Waze have built multibillion-dollar businesses on the ground truth sourced from millions of simultaneous users. Armed with proprietary location technology, they’re poised to preemptively act instead of react, whether by steering drivers toward customers or around congestion.

As ground truth becomes common, it has the potential to transform every industry it touches. Formerly siloed enterprise information—like customer relationship management, business intelligence, and asset management—will be precisely mapped, in real time, and made easily discoverable. Once-hidden spatial patterns and relationships will become visible, with huge ramifications for optimizing supply chains, site selection, and even healthcare during flu season—everywhere big data literally meets the road.

By marrying Mobileye’s sensors with Esri’s industrial-strength mapping and analytics, cities have the opportunity to transform their buses into a rolling fleet of sensors. Not only could they see how, where, and when pedestrians are most at risk, but they could also pass that knowledge back to drivers, warning them of hazards ahead. The data could then be passed on to city planners, police, and anyone else who can help rethink streets to make them safer.

“This data can tell us how and where we move in cities—it’s incredible,” says Sarah Williams, director of the Civic Data Design Lab at MIT. “It can tell us about crashes, potential crashes, great urban design, terrible urban design. And if Esri is the one who visualizes it, so much the better, because everyone who knows mapping knows how to use [their software].”


Hot Spots And Bus Stops

In practice, the data Mobileye collects on Esri’s behalf is prosaic—the time, GPS coordinates, and the directions both the bus and pedestrian were moving at the time of an incident. It’s only when you begin to aggregate, map, and analyze it that useful patterns begin to appear.

In New York City, for example, the kind and frequency of encounters follows the rhythms of the workweek—jaywalking commuters on weekday mornings in Queens, cyclists sidling into blind spots downtown around lunchtime, and school pickups across the five boroughs. “Not only do you see the hot spots, but you also learn why they’re hot and what it actually reveals,” says Nisso Moyal, director of business development and big data at Mobileye.

Directional data yields additional clues. In Manhattan, pedestrians and cyclists tend to flow parallel to the bus, while Brooklynites exhibit perpendicular jaywalking. In some cases, pedestrians stream from subway exits in all directions, a problem with a simple, albeit expensive, solution. “The station is just too close to the road,” says Moyal. “It’s an infrastructure failure that needs to be fixed.” (Easier said than done, of course, especially when the estimated cost to fix New York City’s subways is north of $100 billion.)

The first step in making the infrastructure case is to overlay the history and locations of near misses atop other spatial data sets—an area that Esri, which controls more than half the market for geographic information systems (GIS), excels in. “If you were to aggregate these observations onto the street network and then added the bus stops, you might start to see a relationship between hot spots and the stops,” says Jim Young, Esri’s head of business development. “From a transit planner’s perspective, that might lead you to change the location of a stop.”

The reality for planners is more complex—and the opportunities arguably greater. Stops are already subject to rigorous standards, says Ashley Z. Hand, cofounder of the consulting firm CityFi and former transportation technology strategist for the Los Angeles Department of Transportation. But there are few opportunities for cities to test whether their designs are working as intended.


“One of the challenges cities are facing right now is how they can manage something they can’t measure,” Hand says. Mapping the buses against additional data sets would enable city agencies to inspect bus-stop performance, determine the safest pickup and drop-off points for ride-hailing services, and make the case to an often skeptical public these changes were actually necessary. In New York, local community groups fought Vision Zero improvements to Queens Boulevard, a thoroughfare that has killed 138 pedestrians since 1990, but none since 2014, the program’s first full year. “Data can help drive a better outcome for everyone,” Hand says.

Where The Robots Meet The Road

Buses are only the beginning. Mobileye, Esri, and their counterparts in city halls and transit agencies are all playing a much longer game: autonomous vehicles. “They’re testing their products on a bus so they can use it to train AVs,” argues Williams.

The companies aren’t challenging that assumption. One of the most promising aspects of driverless cars is their potential ability to eliminate up to 90% of auto collisions. Intel purchased Mobileye last year for $15.3 billion in a bid to make its sensors and software the building blocks of an autonomous world. The next iteration of Mobileye’s vehicle-mounted processors “will enable the autonomous-ready city,” Moyal promises. In addition to near-miss collisions, it will recognize lane markings, traffic lights, road signs, and parking spaces, creating a city-owned repository of ground truth for AVs.

With that data in hand, says Esri’s Young, cities will be able to set the rules when it comes to regulating AVs. “Imagine something like a road closure,” he explains. “Rather than give that notice to a specific commercial service, like Waze, the city should publish it to all stakeholders so that it’s useful for everyone,” whether drivers, ride-hailing fleets, emergency services, local businesses, and so on.

In this future, cities won’t be caught flat-footed the way they were with Uber or Lyft. A steady stream of data going from buses to transport operators and city agencies using Esri to AVs will ensure they have the ability to both act and react to changes in real time. Want to alter the flow of traffic through a dangerous intersection? Just refresh the rules for how they should navigate it, backed up by independently verified data.


This is how smart cities will work, as a dynamic organism, adapting continually to ground truth.

This story is created for and commissioned by Esri.


About the author

FastCo Works is Fast Company's branded content studio. Advertisers commission us to consult on projects, as well as to create content and video on their behalf.